Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
American Journal of Energy and Power Engineering 2017; 4(6): 39-43
http://www.aascit.org/journal/ajepe
ISSN: 2375-3897
Keywords Electrical Generation System,
Efficiency,
FFNN
Received: September 21, 2017
Accepted: October 13, 2017
Published: October 25, 2017
High Reliability of Electrical Generation System to Enhance the Efficiency by Using the Algorithm of Feed Forward Neural Networks
Hassan Farhan Rashag*, Mohammed H. Ali
Electronics Technical Department, Technical Institute of Babylon, Al-Furat Al-Awsat Technical
University, Babylon City, Iraq
Email address [email protected] (H. F. Rashag), [email protected] (M. H. Ali) *Corresponding author
Citation Hassan Farhan Rashag, Mohammed H. Ali. High Reliability of Electrical Generation System to
Enhance the Efficiency by Using the Algorithm of Feed Forward Neural Networks. American
Journal of Energy and Power Engineering. Vol. 4, No. 6, 2017, pp. 39-43.
Abstract The electrical generation system is used to generate electrical power based on turbine
and governor. however, the main problem of this classical electrical generation is that the
distortion in the terminal voltage and frequency deviation in the load. To overcome this
problem, new approach of Feed forward neural network FFNN is used to optimize the
performance of system as results to enhance the efficiency of whole system. The
simulation results showed that the system based on FFNN is high reliability and more
effectiveness as compared with traditional system.
1. Introduction
The electrical generation system is widely used to generate high electrical voltage. In
addition, the classical generation system causes high overshoot and undershoot for
current with high oscillation for frequency especially at the starting of load. Therefore,
many authors proposed a lot of method to enhance the performance of system based on
fuzzy logic system [1] [2] [3]. Fuzzy inference system is used instead of PID controller
to enhance voltage [4] [5]. In addition, intelligent techniques us applied on automatic
voltage regulator frequency system to improve the efficiency of system and to minimize
the losses [6] [7].
In this paper, new approach is used to increase the efficiency of electrical system via
FFNN.
2. The Proposed Method Based on Feed Forward
Neural Network
In this proposed method, Simulink matlab is used to build the system using toolbox.
The FFNN is used to optimize frequency deviation and to minimize the distortion of
electrical current. Figure 1 shows the specification of proposed FFNN.
40 Hassan Farhan Rashag and Mohammed H. Ali: High Reliability of Electrical Generation System to Enhance the
Efficiency by Using the Algorithm of Feed Forward Neural Networks
Figure 1. Proposed FFNN.
The regression system for the training, validation, and testing is shown in figure 2
Figure 2. Regression system.
The performance of whole system with training state and FFNN training circuit are showm in figures 3, 4, 5 respectively.
American Journal of Energy and Power Engineering 2017; 4(6): 39-43 41
Figure 3. Performance of FFNN.
Figure 4. Training state of system.
42 Hassan Farhan Rashag and Mohammed H. Ali: High Reliability of Electrical Generation System to Enhance the
Efficiency by Using the Algorithm of Feed Forward Neural Networks
Figure 5. Simulink of FFNN with training.
3. Simulation Results and Discussion
From figure 6, it can be seen that, the behaviour of current under sudden load in the proposed FFNN is more smooth and
free of distortion as compared with traditional method. Figure 7 shows that the efficiency with FFNN is almost 90 percentage
while the efficiency in the classical system 60 percentage.
Figure 6. Compared the electrical current.
American Journal of Energy and Power Engineering 2017; 4(6): 39-43 43
Figure 7. Compared the efficiency of system.
4. Conclusion
The electrical system with FFNN is high accuracy,
robustness and good tracking with the reference. In contrast,
the system with traditional method has low efficiency and
high distortion for current and frequency. Furthermore,
FFNN is used in this paper to remove the problem of
classical system. Finally, the electrical system is enhanced by
using intelligent technique based on Feed forward neural
network.
References
[1] H. L. Demiroren. "the application of artificial neural network technique for AGC" Electric power and energy system" vol. 24. 2002. Pp: 345-354.
[2] H. D. Mathur." frequency stabilization using fuzzy logic for multi area power system". The south pacific journal of natural science. Vol. 4 2007. Pp: 22-30.
[3] Zwee. Lee. "a particle swarm optimization approach for optimum design of PID controller in AVR" IEEE Transaction on energy conversion. Vol. 19. no. 2. pp: 384-392. 2004.
[4] S. Patandi. Design fabrication and parametersevaluation of a surface mounted pmsm. IEEE international conference on power electronics, 2014.
[5] N. B. ulvestad v. kragset. Long step start up simulation for pmsm. 15th international IGTE. 2012.
[6] D, saban. Permanent magnet synchronous machine for subsea application. Technical papers page 1-8 2008.
[7] E, Thibaut. Use of liquid filled motor for pump application. In pcic Europe conference record page 1-8 2010.